**** BEGIN LOGGING AT Fri Sep 14 02:59:56 2007 Sep 14 06:48:52 good morning all Sep 14 06:49:56 morining Sep 14 06:50:08 is there a way to draw a full dependency graph for bootstrap-image? Sep 14 06:50:20 need to know why it wants to build a specific package Sep 14 06:51:38 bitbake -g bootstrap-image is not readable Sep 14 06:51:46 its to small the graph :) Sep 14 06:57:03 use less and grep :-) Sep 14 06:57:22 (seriously, that's how I work out what is requiring what) Sep 14 06:58:18 with the -g option? Sep 14 06:58:21 grep? Sep 14 06:58:53 ah ok got it ... Sep 14 07:02:35 hmm seems that gst-plugins is required by everything Sep 14 07:03:11 im stalling at http://bugs.openembedded.org/show_bug.cgi?id=2382 Sep 14 07:03:14 any1 goit an idea? Sep 14 07:04:18 i use gcc 4.1.2 uclibc 0.9.29 Sep 14 07:45:35 03koen 07org.oe.dev * rcf84b607... 10/ (1 packages/gnome/gconf-dbus_svn.bb): gconf-dbus: sync with poky, drops dependency on gtk+ Sep 14 07:45:39 03koen 07org.oe.dev * rbbdff511... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs: add gconf-dbus Sep 14 07:54:38 gm Sep 14 07:56:57 morning Sep 14 07:59:49 morning all Sep 14 08:07:04 morning Sep 14 08:13:44 hi Sep 14 08:14:04 mrng Sep 14 08:14:58 I'm running into an issue with svn+ssh and bitbake: when bitbake launches svn it doesn't seem to take into account SSH_AGENT_PID and SSH_AUTH_SOCK Sep 14 08:15:25 note that I have no X so I'm launching eval `ssh-agent` and ssh-add manually in my bash shell Sep 14 08:15:59 RP: does pxafb or w100fb make use of CONFIG_FB_CFB_FILLRECT CONFIG_FB_CFB_COPYAREA CONFIG_FB_CFB_IMAGEBLIT CONFIG_FB_DEFERRED_IO CONFIG_FB_MODE_HELPERS options? Sep 14 08:16:55 ok - w100 use few Sep 14 08:21:16 03koen 07org.oe.dev * ra36e2376... 10/ (1 packages/gtk+/gdk-pixbuf-csource-native_2.10.14.bb): gdk-pixbuf-csource-native: add a workaround for #3000 Sep 14 08:23:13 good morning all Sep 14 08:23:34 ~blame koen Sep 14 08:23:35 * ibot blames koen (and Canada) for all the evil in the world Sep 14 08:23:53 hrw: I think they both need them Sep 14 08:24:23 hrw: They're software "accelerated" video routines Sep 14 08:24:27 ok Sep 14 08:24:57 RP: did you had a look at kernel configs which I sent? Sep 14 08:25:10 hi koen Sep 14 08:25:14 hrw: I replied to your mail Sep 14 08:25:29 eeks! Sep 14 08:25:44 * koen looks at http://bugs.webkit.org/show_bug.cgi?id=14998 Sep 14 08:26:35 * florian started to build GTK in and all packages suddenly are $MACHINE arch Sep 14 08:26:53 florian: for gta01? Sep 14 08:27:25 koen: yes Sep 14 08:27:31 right Sep 14 08:27:49 koen: right bug or right feature? Sep 14 08:27:55 mickeyl added the scaling patch only to gta01 and made the gta01 gtk_ MACHINE_ARCH Sep 14 08:28:16 oh Sep 14 08:28:36 Is that patch really only interesting for gta01? Sep 14 08:29:27 it enables lower quality but faster pixbug scaling Sep 14 08:33:54 This might be interesting for some more :) Sep 14 08:44:51 RP: got it. Sep 14 08:46:33 anyone tried --with-softfloat with alsa-libs? Sep 14 08:51:15 any1 aware of a gumstix distributor in europe/germany? Sep 14 09:00:14 * koen flips a coin for "Probability or poking at vendor toolchain patches" Sep 14 09:05:46 hey zecke, where are you now? Sep 14 09:11:14 NOTE: package linux-rp-2.6.22-r9: task do_compile: completed Sep 14 09:11:21 RP: c7x0 builds Sep 14 09:42:55 RP: akita compiles Sep 14 09:44:01 hi Sep 14 09:44:17 I made a library receipt but it didn't install -dev files in $STAGING Sep 14 09:44:46 do someone know where I should look at to solve that ? I'm using autotools Sep 14 09:45:11 do_stage() { autotools_stage_all } into recipe Sep 14 09:45:19 RP: spitz builds Sep 14 09:45:23 hrw: ah ok Sep 14 09:45:24 thanks Sep 14 09:55:17 hrw: The sizes are all ok? Sep 14 09:56:16 * Bernardo built yesterday opie-image/angstrom/akita and angstrom-x11-imag/akita, everything seems ok Sep 14 09:57:20 pidgin 2.2.0 was just released Sep 14 09:57:28 They haven't even announced it :D Sep 14 09:58:18 * koen will wait for 2.2.1 Sep 14 09:58:30 :) Sep 14 09:58:37 Lemme check the changelog first Sep 14 09:58:41 You might not need to Sep 14 09:58:42 RP: yes Sep 14 09:58:55 RP: did you built hx2000 with 2.6.22? Sep 14 10:01:02 changelog: http://pastebin.ca/697301 Sep 14 10:07:27 hrw: Not sure Sep 14 10:10:07 hey Sep 14 10:10:42 koen: I'm still in Eindhoven/Son en Bruegel. Will be in Frankfurt Saturday, Monday, and probably will leave on Tuesday Sep 14 10:10:52 ok Sep 14 10:11:08 zecke: you're very close to both likewise and gerwinin Sep 14 10:11:35 koen: so from Tuesday/Wednesday on we could meet some where. Amsterdam would be my preferred_provider :) Sep 14 10:12:08 I'm trying to fit you into my schedule :) Sep 14 10:12:14 zecke: you travel from Eindhoven to Frankfurt today or tomorrow? Sep 14 10:12:20 RP: poodl built Sep 14 10:12:31 * koen heads out for lunch Sep 14 10:12:34 later all Sep 14 10:12:40 need ro get some food... Sep 14 10:13:26 RP: so poodle, c7x0, akita, spitz builds. hx2000 fails, tosa is in bad shape so I ignore it (same for collie) Sep 14 10:14:05 hrw: This was with both jffs2 and ext3 enabled? Sep 14 10:14:17 no Sep 14 10:14:28 jffs2 for flashboot, ext3 for hddboot Sep 14 10:15:06 tosa needs a whole lot of TLC Sep 14 10:15:44 hrw: ok, that makes sense Sep 14 10:27:27 hi mickeyl Sep 14 10:27:51 morning Sep 14 10:27:52 hi hrw Sep 14 10:28:17 zecke: are you ok with Lasagne on Sunday evening or is that not good for your stomach? Sep 14 10:28:38 RP: tcuniversal builds Sep 14 10:29:12 mickeyl: that is fine Sep 14 10:29:36 zecke: great. We'll get Lasagine a Sabine then Sep 14 10:29:42 s/Lasagine/Lasagne/ Sep 14 10:30:30 RP: sent update to kernel config split Sep 14 10:30:47 mickeyl: vegetarian? just kidding Sep 14 10:31:16 nah, meat meat meat Sep 14 11:20:47 gm Sep 14 11:21:03 With the gumstix, I have the kernel module for the network interface insalled Sep 14 11:21:10 but it does not autoload Sep 14 11:21:31 one problem is, I do not know in advance if the network card is intalled Sep 14 11:21:51 is there a way to have Angstrom probe for hw and load modules at run time? Sep 14 11:22:41 Crofton|home: what are the downsides of loading both drivers? Sep 14 11:22:55 an extra 8 or 12k or ram used isn't a downside for me Sep 14 11:23:09 if the card is not there you will get an annoying boot message? Sep 14 11:23:15 ooooh Sep 14 11:23:19 I'm scared Sep 14 11:23:26 of the bug reports :) Sep 14 11:23:49 afaik you won't get a message Sep 14 11:23:56 the module should silently load Sep 14 11:23:58 so I could use the modules-autoload thing in a bb file? Sep 14 11:24:23 I have that on the OSK because I am lazy Sep 14 11:24:57 Aren't modules supposed to autoload when needed somehow? Sep 14 11:25:48 module_autoload_ohci-hcd_omap5912osk = "ohci-hcd" Sep 14 11:25:56 so use a line like this in linux.inc? Sep 14 11:28:59 pretty much Sep 14 11:30:13 * koen|away heads home Sep 14 11:31:10 any chance it could go in the machine file? Sep 14 11:31:18 koen: thx for the link to the n800 schematics on your site. Sep 14 11:37:22 Crofton|home: how that network card is connected? Sep 14 11:37:41 hrw, you are familiar with gumstix? Sep 14 11:37:48 there is a processor board Sep 14 11:37:56 and a board with the network interface Sep 14 11:37:58 Crofton|home: cf? spi? sdio? Sep 14 11:38:11 so you can run the processor with or without an attached network card Sep 14 11:38:17 urgh Sep 14 11:38:25 that I would have to research Sep 14 11:38:36 Crofton|home: I know - but maybe udev hotplug can detect that this card is connected Sep 14 11:38:49 it does not at the moment Sep 14 11:39:10 Craig from gumstix has been mia for a day or so Sep 14 11:39:22 I want to ask him how he handles this in the build root system they use Sep 14 11:39:38 but, I need some stuff to work for a demo next week :) Sep 14 11:39:47 in buildroot they choose 'gumstix with network attached' Sep 14 11:39:51 probably Sep 14 11:39:53 ah Sep 14 11:39:58 heh Sep 14 11:40:23 gumstix will be a nuisance due to the availabel hw permuations :) Sep 14 11:41:35 hrw, can the module_autoload be in the machine file, or does it need to be in linux.inc? Sep 14 11:41:50 it can be in both Sep 14 11:41:53 ok Sep 14 11:42:06 I prefer to keep it with the machine and not clutter linux.inc Sep 14 11:42:36 My principal concern with linux.inc is managing clutter so people are not intimidated by it Sep 14 11:42:45 re Sep 14 11:43:00 I also prefer the principal of limiting breakage by mistakes Sep 14 11:43:00 Crofton|home: that's cosmetices Sep 14 11:43:06 yeah Sep 14 11:43:13 but cosmetics are important Sep 14 11:43:22 there aren't :) Sep 14 11:43:27 only for PHBs Sep 14 11:43:33 ah.. I will give koen's autobuild hard task with push.. Sep 14 11:43:58 * Crofton|home hopes koen becomes a PHB one day :) Sep 14 11:44:36 koean cosmetics are also for old guys who cannot keep a large number of things in their head at one time :) Sep 14 11:45:22 "putting things in non-obvious places to avoid clutter" isn't on my list of great development tactics Sep 14 11:45:54 modules that should load on a per machine basis belong with the machine Sep 14 11:46:25 The modules_autoload line would already be tagged as machine specific Sep 14 11:46:30 you need to rebuild the kernel to have to take effect..... Sep 14 11:46:35 that I know Sep 14 11:46:49 so putting it in machine.conf is misleading Sep 14 11:46:51 that is an interesting viewpoint though Sep 14 11:47:45 so there is a functional relationship (machine) and a build time relationship (the bb file that must be rebuilt) Sep 14 11:48:03 it depends on how you view the meta-data Sep 14 11:48:40 The problem is that it works in both places :) Sep 14 11:49:16 the t2 system puts all machine and arch related patches in the machine and arch subdirs Sep 14 11:50:43 this is the way I think :) Sep 14 11:51:03 I understand grouping stuff on a build basis also Sep 14 12:07:02 can i somehow define this on the command line for one bitbake invocation: do_build[recrdeptask] = "do_build" Sep 14 12:13:10 tuukkah: wants to replicate what meta.bbclass does, but on a package basis. I advised him to just build a feed, but he is interested to know if it can be done for a single package. Sep 14 12:14:59 'bitbake single-package' will run all tasks Sep 14 12:16:14 koen, but i think i want an added do_build[recrdeptask] = "do_build" Sep 14 12:16:44 'bitbake single-package' will run all tasks Sep 14 12:17:06 koen: does it build ipkgs for all RDEPENDS? Sep 14 12:17:07 ok so you think i don't want that. thanks Sep 14 12:17:35 rwhitby: it won't, but that wasn't the question Sep 14 12:17:53 tuukkah: koen likes to play games giving answers - he will not give you the right answer until you ask the question in exactly the correct way Sep 14 12:18:01 koen, you didn't really answer my question either now did you? Sep 14 12:18:14 I did Sep 14 12:18:24 tuukkah: and there is no point in getting angry with him about it, it's just the way he is. Sep 14 12:18:26 rwithby said "single package" Sep 14 12:18:42 see? :-) Sep 14 12:18:51 i asked "can i" so the answer would be "no" or how, or "yes" if you don't want to communicate Sep 14 12:19:15 tuukkah, on irc you have to accept people for themselves Sep 14 12:19:30 and ask proper questions Sep 14 12:19:36 since my crystal ball is in repairs Sep 14 12:19:47 Crofton|home, i don't have to accept anything, i can always leave you in peace Sep 14 12:19:52 you cannot depend on them to be what you want them to be :) Sep 14 12:19:54 we also develop very thick skins and can be very terse Sep 14 12:20:00 tuukkah, exactly :) Sep 14 12:20:30 tuukkah: and don't take any of it personally, koen is like that with everyone. Sep 14 12:20:49 (except those who ask the questions in precisely the right way) Sep 14 12:21:37 you also learn to read what he writes, because there is usually somethign useful in the text Sep 14 12:22:08 yeah, it's often like a cryptic crossword puzzle - the answer will be in there, but you have to really work to find it ;-) Sep 14 12:24:01 you guys need to interact with the Dutch more :-) Sep 14 12:24:18 tuukkah: so the decoding of koen's answer is that ipkgs for REDENDS are not built by default when you bitbake a single package. Sep 14 12:24:27 RDEPENDS Sep 14 12:25:01 as I understand it, they are only built if that package is part of an image that you build, or part of another package which includes meta.bbclass (or equivalent) Sep 14 12:25:16 i.e. the normal stuff you do for images or feeds. Sep 14 12:25:47 it also depends on the age of your metadata Sep 14 12:26:32 two weeks ago it would write out stuff, but since we have do_package_write_ipk and do_package_write_deb the packages aren't getting generated by default Sep 14 12:26:41 unless you use -c rmall and friends Sep 14 12:27:27 is that an additional condition on top of the ones I stated, or an alternative condition? Sep 14 12:28:51 XorA: is it really a Dutch thing? I must admit that I haven't interacted with many Dutch, so I wouldn't know if it's a generalisation of Dutch engineers or just Koen :-) Sep 14 12:29:08 it's a dutch thing Sep 14 12:29:53 it's funny, but that somehow makes it easier for me to accept it :-) Sep 14 12:30:42 "standing in line" and "waiting for your turn" are also unknown things in .nl as well Sep 14 12:30:44 rwhitby: yeah, I have another Dutch friend, exactly the same communication style as koen, thats why I dont notice it, I spent 5 years working with him Sep 14 12:32:16 re Sep 14 12:32:50 I wonder if RP knows it's a Dutch thing. Sep 14 12:42:14 Anyone trying to use AUTOREV should see the mail I just sent to oe-devel. bitbake is broken with it atm :( Sep 14 12:42:47 * Darth_Wader hits koen with a poisoned sledgehammer Sep 14 12:43:43 RP: good thing I removed it from openmoko.conf ;-) Sep 14 12:43:54 Darth_Wader: what for? Sep 14 12:44:07 for h2200 leds Sep 14 12:44:18 Darth_Wader: I added a defconfig that worked completely till paul screwed it up Sep 14 12:44:43 i`m not sure whose fault it is, but leds stuff afaiu can`t be fucked up by kernel config Sep 14 12:44:59 [ 2062.160000] hamcop_set_led: illegal LED number 5, must be < 5 Sep 14 12:45:28 rwhitby: It may well be a dutch thing but I'm sure koen can learn to communicate more clearly ;-) Sep 14 12:46:31 RP: I get the feeling that it's a like a badge of honour with koen to make the questioner ask in precisely the correct way. I believe that koen knows exactly how to communicate clearly, but sometimes chooses not to. Sep 14 12:47:01 rwhitby: I know, which is what irritates me at times ;-) Sep 14 12:47:27 me too, but now that I know how to recognise it instantly, I can work around it. Sep 14 12:47:29 "you get what you ask for" instead of "you get what you need"? Sep 14 12:47:57 yep, that's a good summary Sep 14 12:48:08 We should add that to the topic as a warning :) Sep 14 12:48:52 speaking of topic ;D Sep 14 12:49:30 bah, who would have time for word plays with strangers over irc Sep 14 12:49:56 koen: As with anything, things can go to an extreme and whilst any regulars here know what to expect, it is a little hard on newcommers ;-) Sep 14 12:50:33 ~smart questions Sep 14 12:50:34 * ibot plops a heatsink on questions's head so questions's brain so it doesn't overheat. Sep 14 12:50:49 ~smartquestions Sep 14 12:50:50 smartquestions is probably http://www.catb.org/~esr/faqs/smart-questions.html Sep 14 12:58:12 from my igoogle page Sep 14 12:58:15 http://www.quotationspage.com/quote/27543.html Sep 14 12:58:54 stupidquestions is http://www.quotationspage.com/quote/27543.html Sep 14 12:58:55 Crofton|work: I read the log, so seamonkey is on holidays, right? Sep 14 12:59:02 ~stupidquestions Sep 14 12:59:05 yeah Sep 14 12:59:25 apparently his missus thinks a trip to france for a wedding should last 10 days ... Sep 14 12:59:34 spacemonkey actually Sep 14 12:59:36 :> Sep 14 12:59:42 oh k, soz Sep 14 12:59:51 :) Sep 14 12:59:57 we need to figure out why the git kernle does not boot it seems Sep 14 13:00:12 well, we know its because of the board type mismatch thing Sep 14 13:00:20 question is why that happens Sep 14 13:00:24 do we? Or do we just suspect? Sep 14 13:00:36 Crofton|work: I enabled some debug options that koen suggested Sep 14 13:00:40 and I got the output saying that Sep 14 13:00:42 have you instelled git u-boot? Sep 14 13:00:45 ok Sep 14 13:00:50 excellent Sep 14 13:00:51 no, not yet Sep 14 13:00:57 I have not yet figured out how to reflash uboot :) Sep 14 13:01:00 maybe that will help Sep 14 13:01:10 you need dvfalsher Sep 14 13:01:18 er dvflasher Sep 14 13:01:30 from TI website Sep 14 13:01:48 couldn't I just use ubot to flash itself? Sep 14 13:01:56 you could do that also Sep 14 13:02:12 it would be good to know how to use dvfalsher in case you screw up :) Sep 14 13:02:25 :) Sep 14 13:02:28 indeed Sep 14 13:03:35 http://wiki.davincidsp.com/index.php?title=DVEVM_and_DVSDK_frequently_asked_questions Sep 14 13:03:41 bbiab Sep 14 13:03:49 thx Sep 14 13:03:59 ibot: stupidquestions is http://www.quotationspage.com/quote/27543.html Sep 14 13:04:00 okay, rwhitby Sep 14 13:04:08 ~stupidquestions Sep 14 13:04:09 hmm... stupidquestions is http://www.quotationspage.com/quote/27543.html Sep 14 13:05:04 03hrw 07org.oe.dev * r80e71d36... 10/ (1 conf/distro/include/moko-autorev.inc): moko-autorev: add also libraries: libmoko* and libgsmd Sep 14 13:05:10 03hrw 07org.oe.dev * r126555b0... 10/ (1 conf/machine/include/tune-geodelx.conf): Sep 14 13:05:10 tune-geodelx: added tuning for AMD GeodeLX cpu Sep 14 13:05:10 Starting from GCC 4.3 it will have own -mcpu flag. With older versions Sep 14 13:05:10 optimizations for AMD K6-2 are best. Sep 14 13:05:16 03hrw 07org.oe.dev * rd92d2182... 10/ (1 conf/machine/alix.conf): Sep 14 13:05:16 alix: use tuning for GeodeLX cpu, switch to i586 Sep 14 13:05:16 According to Jordan Crouse (Geode maintainer in Linux) GeodeLX is i586 not Sep 14 13:05:17 i686. Also optimizations for AMD K6-2 are due to his suggestions. Sep 14 13:05:19 My ALIX.1c board runs with Geode optimized packages for over week without Sep 14 13:05:21 problems. Sep 14 13:12:58 good morning Sep 14 13:13:32 Crofton: did you get your gumstix question answered? Sep 14 13:23:00 no :) Sep 14 13:23:52 modprobe smc91x did startup the network Sep 14 13:23:59 so I will try autoloading it Sep 14 13:26:16 ~seen hopsnbarley Sep 14 13:26:19 hopsnbarley was last seen on IRC in channel #oe, 2d 21h 40m 1s ago, saying: 'hi likewise!'. Sep 14 13:28:28 hmm, I have missed that :-) Sep 14 13:28:38 :) Sep 14 13:31:04 03mickeyl 07org.oe.dev * r8caa6e70... 10/ (1 packages/vte/vte_0.16.8.bb): vte: add rdependency on termcap Sep 14 13:31:16 03mickeyl 07org.oe.dev * r838a60f4... 10/ (1 packages/termcap packages/termcap/termcap_11.0.1.bb): add termcap, the terminal capabilities file. at least vte needs this to operate correctly (backspace et. al.) Sep 14 13:31:28 03koen 07org.oe.dev * rcaed6636... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs: bump libgsmd Sep 14 13:31:40 03koen 07org.oe.dev * r4b8f7f26... 10/ (8 files in 4 dirs): linux-ezx: improve rokre6 support Sep 14 13:39:38 03koen 07org.oe.dev * rfbf6625a... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs: bump openmoko-icons2 Sep 14 13:43:03 hmm devshell doesn't set the PATH to staging/i686 ? Sep 14 13:44:27 ~curse gsmd Sep 14 13:44:28 May you be reincarnated as a Windows XP administrator, gsmd ! Sep 14 13:47:59 koen: building our rootfs, due to pb to access git trees, I am stuck on mtd-utils building (mtd-utils_1.0.0+git.bb) ; anyway, I saw that there is some mtd-utils git archives stored in angstrom source server, how can I use them ? Sep 14 13:53:54 hoo yaa! booted angstrom on my storcenter :> Sep 14 13:54:35 storcenter? Sep 14 13:54:52 Crofton|work: Iomega StorCenter, a NAS thingie Sep 14 13:54:56 ah Sep 14 13:54:57 cool Sep 14 13:55:01 PPC board Sep 14 13:55:28 hmm, actually an angstrom-nas image would be nice Sep 14 13:56:56 btw what does that mean? -sh: can't access tty; job control turned off Sep 14 13:57:09 I started with the angstrom minimal image Sep 14 14:00:27 mickeyl: It should do... Sep 14 14:02:46 RP: ah right. it gets overridden in my .bashrc Sep 14 14:02:59 how do you differenciate a shell window from a devshell window? Sep 14 14:03:17 mickeyl: Good question Sep 14 14:03:28 hmm i guess we could set a variable Sep 14 14:03:33 and then i can check for that in my .bashrc Sep 14 14:04:06 mickeyl: You have all kinds of bitbake variables exported in the devshell, there is probably on you could use Sep 14 14:04:30 yeah, don't need to invent a new one Sep 14 14:04:37 mickeyl: I thought you mean how to tell just looking at the shell prompt :) Sep 14 14:05:10 heh, no it's really for the .bashrc since i have lots of initialization there Sep 14 14:09:26 Crofton: sorry had to step away Sep 14 14:09:37 np Sep 14 14:09:50 Yes, autoloading is the first step Sep 14 14:10:22 do you know why there is a smc91x module and a gumstix-smc91x module? Sep 14 14:11:15 No, not for certain. Perhaps something to do with setting up GPIO's for the smc hw? Sep 14 14:11:49 Never bothered to investigate since I use the wifi module for networking Sep 14 14:12:08 Crofton: if its anything like the cs89xx driver its because network drivers dont follow the platfrom driver model and there is no way for the board file to pass resources Sep 14 14:12:09 Well, I am working on flashing an image that should try and load module automatically Sep 14 14:12:34 I am guessing the smc91x module depends on it so it is loaded? Sep 14 14:12:40 I do not have a machine booted atm Sep 14 14:12:55 Crofton|work: modinfo smc91x Sep 14 14:14:14 having a strange issue with my sdr stuff atm I need to fix Sep 14 14:15:10 root@sakoman-gumstix-custom:~$ modinfo smc91x Sep 14 14:15:10 filename: /lib/modules/2.6.21/kernel/drivers/net/smc91x.ko Sep 14 14:15:10 license: GPL Sep 14 14:15:10 depends: mii,gumstix-smc91x Sep 14 14:15:10 vermagic: 2.6.21 preempt mod_unload ARMv5 Sep 14 14:15:11 parm: nowait:set to 1 for no wait state (int) Sep 14 14:15:13 parm: watchdog:transmit timeout in milliseconds (int) Sep 14 14:15:28 ok, so load smc91x and it works Sep 14 14:16:22 Crofton: you will also want to overide the default /etc/network/interfaces in netbase to add: "auto eth0" Sep 14 14:16:51 Then all should be well Sep 14 14:19:38 03crofton 07org.oe.dev * refe3244e... 10/ (1 packages/exmap-console/exmap-console.inc): exmap-console : Add dependency on glib-2.0. Sep 14 14:31:00 koen: ok, found how to deal with this. Sep 14 14:31:37 rp, who is the guy doing exmap-console in Poky? Sep 14 14:31:44 hi, are there any chanches to get a comment on oe bug 2598. would be nice to know if it has chances to get into oe. Sep 14 14:32:17 Crofton: Tomas Frydrych (tf@oh) Sep 14 14:32:28 oh dot com? Sep 14 14:32:31 I also asked about userland or kernelspace on the arm maling list and it sounds that userland is prefered Sep 14 14:32:36 Crofton: o-hand.com Sep 14 14:32:40 I had to add a dependency on glib-2.0 Sep 14 14:32:42 ah Sep 14 14:32:47 I always mess that up Sep 14 15:02:18 how do i use a custom gcc in my local.conf? I can't find the wiki page with that info Sep 14 15:06:09 Hi, I'm trying to track down a boot problem with OE on a ep93xx board. My kernel boots then gives me "Warning: unable to open an initial console", afaict this is a lack of /dev/console in the rootfs. In the OE rootfs directory there appear to be no devs, are these made as part of the build or are they created at runtime? Sep 14 15:09:08 created by udev I think Sep 14 15:09:20 I'll add the udev package then :) Sep 14 15:09:34 what distro are you using? Sep 14 15:09:35 I think you can also have a device table file? Sep 14 15:09:44 Crofton: a hacked slugos. Sep 14 15:09:48 hmmm Sep 14 15:10:02 I do not know what the correct answer is for you .... Sep 14 15:10:12 trj: on runtime by /etc/init.d/makedevs iirc Sep 14 15:10:44 Is RDEPENDS_kernel-base = "" the correct way to keep the kernel image out of the root fs Sep 14 15:11:03 hrw: ah, my rootfs doesn't have the makedevs init script, ta. Sep 14 15:11:05 * Crofton is clearing up work he has forgotten about since last week Sep 14 15:34:25 morning Sep 14 15:36:35 03mickeyl 07org.oe.dev * r9809cb68... 10/ (1 packages/ttf-fonts/ttf-liberation_0.2.bb): add ttf-liberation fonts Sep 14 15:36:40 03mickeyl 07org.oe.dev * r7bb17579... 10/ (1 packages/openmoko2/openmoko-terminal2_svn.bb): openmoko-terminal2: add rdependency on ttf-liberation-mono Sep 14 15:52:08 mickey|sports: Poky has ttf-liberation latest Sep 14 15:53:04 bye all Sep 14 15:54:29 is anyone here using swap partitions with their kernel? I have a nand flash with plenty of room and was trying to see if it made a difference in speed Sep 14 15:57:07 HopsNBarley: hey, do you have a minute? Sep 14 15:57:37 sure! Sep 14 15:57:49 HopsNBarley: are you somehow involved in the openprotiumthing? Sep 14 15:58:13 yes i am. Sep 14 15:58:28 you guys have your own tree, right? not in OE? Sep 14 15:58:42 all our work is in published in OE. Sep 14 15:58:53 so you sync from time to time? Sep 14 15:59:00 reason I am asking: I got a 1TB Quad Sep 14 15:59:37 sadly, i'm a few months behind due to a tough consulting job, but that is behind me now, so it'll now move forward. Sep 14 15:59:45 mickey|sports: the EFL cvs recipes you added have a lower PV than the releases you removed :( Sep 14 15:59:45 oh, nice. I can help you flash that. Sep 14 15:59:57 HopsNBarley: I already have angstrom minimal image running on it Sep 14 16:00:05 openprotium did not build for me for some reason Sep 14 16:00:19 HopsNBarley: did you change anything in the kernel yourself, or did you simply apply the Iomega patches? Sep 14 16:00:20 mtn: fatal: std::logic_error: roster.cc:188: invariant 'fetching nonexistent entry from children' violated Sep 14 16:00:24 mtn: this is almost certainly a bug in monotone. Sep 14 16:00:25 wtf? Sep 14 16:00:28 Jin^eLD, cool. oh, too bad. well, we'll get that working soon. Sep 14 16:00:50 maybe we should not floot the channel with storcenter specifics, is privmsg ok? Sep 14 16:01:00 Jin^eLD, we are somewhere in the middle of getting a single kernel for both. sure, go priv. Sep 14 16:17:24 guys, any idea? trying to build angstrom-console-image Sep 14 16:17:26 http://pastebin.ca/697632 Sep 14 16:17:31 why is there this ARCH mismatch? Sep 14 16:18:23 Jin^eLD: powerppc/ppc is being overhauled in the kernel Sep 14 16:18:47 so depending on your kernel version arch is either powerpc or ppc Sep 14 16:18:50 overhauled you mean overruled by some setting? Sep 14 16:18:53 aha Sep 14 16:18:58 Jin^eLD: ask steliosk for the details Sep 14 16:19:04 oh ok, thanks Sep 14 16:19:22 I don't understand it either, I just know it's trouble :( Sep 14 16:20:12 03koen 07org.oe.dev * r24425de4... 10/ (1 packages/efl1/ecore.inc packages/efl1/edje_cvs.bb): ecore, edje: try to repair PV going backwards Sep 14 16:22:21 03florian 07org.oe.dev * rf9da0188... 10/ (3 files in 3 dirs): gpe-session-scripts-phone: Minor session update to avoid session restart. Sep 14 16:22:37 03florian 07org.oe.dev * rc8ff637d... 10/ (3 files in 2 dirs): gpe-applauncher: Make a separate package for the machine specific configuration files. Sep 14 16:22:43 03florian 07org.oe.dev * r9efbf30d... 10/ (1 packages/tasks/task-sdk-extra.bb): task-sdk-extra: Some updates, add libraries, drop obsolete. Sep 14 16:22:50 03florian 07org.oe.dev * ra1f6076d... 10/ (1 packages/gpephone/gpe-applauncher_0.11.bb): gpe-applauncher: add 0.11 Sep 14 16:22:54 03florian 07org.oe.dev * r5fab22ef... 10/ (8 files in 4 dirs): gpe-applauncher: add svn recipe, ship device specific softkey configuration. Sep 14 16:23:00 03koen 07org.oe.dev * r24425de4... 10/ (1 packages/efl1/ecore.inc packages/efl1/edje_cvs.bb): ecore, edje: try to repair PV going backwards Sep 14 16:26:28 steliosk: ping Sep 14 17:00:05 mhm, I think I screwed up the ppc/powerpc thing in my kernel bb file... I did override and set the ARCH there; but when I don't, when the arch is "powerpc" the kernel will not build Sep 14 17:00:16 linux-2.6.12.6/arch/powerpc/Makefile: No such file or directory Sep 14 17:00:30 because it is looking in "powerpc" but in the kernel tree (old stuff) its still "ppc" Sep 14 17:16:37 Jin^eLD : which cpu ? Sep 14 17:17:43 steliosk: I think I figured it out, its ppc603e that uses a 2.6.12 kernel Sep 14 17:17:58 madwifi has overrides for that, but not for my machine Sep 14 17:18:06 so the prob was actually there and not in the kernel bb Sep 14 17:28:03 Jin^eLD : i know. I have added them :) Sep 14 17:28:15 steliosk :)) Sep 14 17:28:39 Jin^eLD : ppc603e is efika also Sep 14 17:28:43 btw, what do you guys think about an angstrom-nas-image? would be a nice thing to have... Sep 14 17:28:55 ah, I see Sep 14 17:29:01 Jin^eLD : so it should build without problems Sep 14 17:29:11 Jin^eLD : what machine is that ? Sep 14 17:29:16 a new one Sep 14 17:29:20 based on storcenter Sep 14 17:29:27 that's why it was not in already Sep 14 17:29:32 and that's why I ran into the problems :) Sep 14 17:29:52 but makes sense now Sep 14 17:31:00 Jin^eLD : if your cpu is still in the ppc tree (i.e 405/440 etc) you will also need to copy some headers during the kernel build Sep 14 17:31:48 afaik the exact cpu is "ppc603e" Sep 14 17:32:05 so I am fine then, right? Sep 14 17:32:57 Jin^eLD : yep Sep 14 17:35:24 doh.. madwifi won't compil Sep 14 17:35:25 e Sep 14 17:35:27 I must be missing something Sep 14 17:47:02 what does that notation mean? ${@base_contains("MACHINE_FEATURES", "screen", "psplash-zap", "",d)} Sep 14 18:02:48 does anyone know, what's the minimum kernel version for madwifi-ng? Sep 14 18:03:19 madwifi.org? Sep 14 18:03:50 good hint.. Sep 14 18:12:09 Was that problem with udev/makedevs ever fixed? Sep 14 18:13:39 Where you don't get static /dev/ entries (essential stuff, like /dev/null) unless you uncomment /dev in device_table.txt of package initscripts (which is not a proper solution). Sep 14 18:14:08 I'm getting some errors at the "udev is populating /dev" step Sep 14 18:14:12 At bootup Sep 14 18:14:29 Missing files, general brokeness :) Sep 14 18:35:11 steliosk, ping Sep 14 18:35:25 TheSheriif, I think you mispelled your nick .... Sep 14 18:35:31 yup Sep 14 18:35:36 fixable Sep 14 18:35:49 now we wait for steliosk to notice Sep 14 18:35:53 03koen 07org.oe.dev * r2f6b0d44... 10/ (5 files in 2 dirs): efl: more PV unbreaking Sep 14 18:38:11 TheSheriif, go to www.openembedded.org and go through getting started Sep 14 18:41:10 thats better Sep 14 18:41:26 hmmm is it me or 2.6.22 is a pile of %#^#$$&$&%&%^&*^* Sep 14 18:42:36 chouimat|ibook: your statements not mutually exclusive :) so it may be both Sep 14 18:42:53 :) Sep 14 18:53:48 if I add RDEPENDS_kernel-base to a machine file, what do I need to rebuild? Sep 14 19:00:49 Crofton : pong Sep 14 19:00:51 hey Sep 14 19:00:54 TheSheriff, ping Sep 14 19:01:09 pong Sep 14 19:01:14 TheSheriff, steliosk is the guy that knows how to make Linux for the ml403 Sep 14 19:01:17 does this look familiar to anyone? -sh: can't access tty; job control turned off Sep 14 19:01:29 excellent Sep 14 19:01:36 Jin^eLD, I think that is a missing entry for a console in /dev? Sep 14 19:01:43 * steliosk ducks Sep 14 19:01:48 heh Sep 14 19:01:51 haha Sep 14 19:01:54 Crofton: I have a console there... Sep 14 19:01:54 that bad? Sep 14 19:01:55 I am helping him get OR going Sep 14 19:02:02 so must be something else Sep 14 19:02:02 Jin^eLD, that was my guess Sep 14 19:02:08 google for the message? Sep 14 19:02:27 steliosk, so after he gets OE going he may have some questions Sep 14 19:02:31 doing that right now, loads of stuff, I was hoping someone encoutnered that with angstrom Sep 14 19:02:34 and new the answer :) Sep 14 19:02:35 I will loan him my board Sep 14 19:02:40 yeah Sep 14 19:02:44 I know what you mean Sep 14 19:04:14 Crofton TheSheriff : your board is a bit different than ml403 but it will be good chance to add that also to OE Sep 14 19:04:34 I have a 403 :) Sep 14 19:04:46 the xilinx eval one Sep 14 19:05:25 Crofton TheSheriff : a cool then. i though you had a XUP or somthing like that with the VIrtex pro Sep 14 19:05:37 not yet :) Sep 14 19:05:55 I bought the ml403 because I knew you were working on it Sep 14 19:06:05 and it has the peripherals we need Sep 14 19:06:10 uuh.. the message is because the shell is on /dev/console and not on /dev/tty Sep 14 19:06:23 have you had much luck with the ethernet? Sep 14 19:06:27 why would angstrom pick /dev/console? or where is that set at all heh.. Sep 14 19:06:43 * steliosk calls xilinx to get his commission Sep 14 19:07:41 Crofron : Yes. We have a bitstream with 10/100/1000 ethernet,uartlite,ps2,buttons,leds etc Sep 14 19:07:55 Crofron : and runs asterisk fine :) Sep 14 19:08:06 great Sep 14 19:08:20 * Jin^eLD screwed up as always Sep 14 19:09:27 I'll get him started and get him to where he asks questions I cannot answer Sep 14 19:09:31 Crofron : working on getting vga support added Sep 14 19:09:37 well, I need to do my homework first and figure out how this all works Sep 14 19:09:43 good Sep 14 19:09:46 no flunking Sep 14 19:10:02 on it Sep 14 19:10:09 he needs to learn to start a build, then do homework :) Sep 14 19:11:42 so let me get this straight Sep 14 19:12:24 OpenEmbedded builds an image Sep 14 19:12:40 then I can use the Xilinx tool to add support for peripherals Sep 14 19:12:58 Crofton: lol. Yes, during the build there is enough time for homework. Or a thesis even ;-) Sep 14 19:13:12 heh Sep 14 19:13:23 sdr forum conference paper atm Sep 14 19:13:37 Crofton: nice. conference where? Sep 14 19:13:43 Denver Sep 14 19:13:56 a bit early for skiing, but some resorts should be open Sep 14 19:17:44 TheSheriff : Nop. first you design your fpga. Add core,peripherals etc. Then you generate the BSP (from withing the xilinx enviroment) Sep 14 19:18:06 TheSheriff : Then you use OE Sep 14 19:19:29 because BSP output flows into kernel, right? Sep 14 19:20:02 yes Sep 14 19:21:54 time to watch basketball Sep 14 19:22:09 * Crofton wonders how partial reconfiguration works in this sort of design Sep 14 19:22:28 i don't think it does Sep 14 19:22:42 erm Sep 14 19:22:42 [ 74.930000] JFFS2 notice: (326) check_node_data: wrong data CRC in data node at 0x003e2a00: read 0x8f9d2157, calculated 0x5b6abb17. Sep 14 19:22:51 is this sign of broken flash? Sep 14 19:23:43 Darth_Wader: probably means you had a hard shutdown: http://www.linux-mtd.infradead.org/archive/tech/faq.html Sep 14 19:24:34 hmm Sep 14 19:24:46 Darth_Wader: I've never see a JFFS2 fs go bad because of these messages -- come to think of it, I've not ever seen a jffs2 fs go bad in recent history ... Sep 14 19:28:39 bye everyone Sep 14 19:28:46 thanks for the help Sep 14 19:34:31 DISCON Sep 14 19:34:34 hi khem Sep 14 19:34:47 likewise: howdy Sep 14 19:36:30 khem: newest eglibc doesn't need the single patch we have in OE I saw. However, it doesn't build for me anymore. I have stuck with 3447 for the moment. Sep 14 19:36:56 that's right we pulled changed from upstream Sep 14 19:49:34 Darth_Wader: that seems to happen all of the time for me. heck, I have that, and messages about empty parts when I do a "reboot" command. Sep 14 20:01:33 zecke: hi Sep 14 20:01:43 hi zecke Sep 14 20:55:14 re Sep 14 21:58:40 03mickeyl 07org.oe.dev * r51a3cb65... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs: bump openmoko-terminal2 Sep 14 21:59:05 mickeyl: Did you see my post to oe-devel about SRCREV? Basically bitbake needs a way of knowing when SRCREV uses the get_srcrev function and when its set to a fixed value. Any ideas how to do this generically? Sep 14 22:00:19 seen it, but had no time to think about it yet Sep 14 22:30:25 mickeyl: looks like sane-srcrevs bumping is working so far ;-) Sep 14 22:31:03 Just need to track down the last date based recipes now ... Sep 14 22:35:23 rwhitby: lets see ;) Sep 14 22:36:11 yeah, it's still honeymoon period Sep 14 23:04:02 03rwhitby 07org.oe.dev * r0a9f655b... 10/ (5 files in 3 dirs): openocd: Converted to SRCREV and added to *-*revs?.inc Sep 14 23:05:29 03rwhitby 07org.oe.dev * r0a9f655b... 10/ (5 files in 3 dirs): openocd: Converted to SRCREV and added to *-*revs?.inc Sep 14 23:11:32 hmm, anyone seen such errors? Sep 14 23:11:37 NOTE: package gpe-mini-browser-0.21-r0: task do_configure: failed Sep 14 23:11:45 other here: http://pastebin.ca/697583 **** ENDING LOGGING AT Sat Sep 15 02:59:57 2007